Re: [PATCH v1 3/7] powercap/dtpm: Fixup kfree for virtual node

[Daniel Lezcano]

On 22/02/2022 16:55, Ulf Hansson wrote:
> On Fri, 18 Feb 2022 at 14:18, Daniel Lezcano <daniel.lezcano@xxxxxxxxxx> wrote:
> > On 17/02/2022 16:45, Ulf Hansson wrote:
> >
> > [ ... ]
> >
> > > Does ops->release() also resets the "dtpm" pointer to NULL? If not,
> > > it's good practice that it should, right?
> > >
> > > In that case, we would be calling "kfree(NULL);" the second time,
> > > which is perfectly fine.
> >
> > So you suggest to replace:
> >
> > if (ops->release)
> >          ops->release(dtpm);
> > else
> >          kfree(dtpm);
> >
> > By:
> >
> > if (ops->release) {
> >          ops->release(dtpm);
> >          dtpm = NULL;
> > }
>
> I don't have a strong opinion how to code this.
>
> What I was trying to point out was that if ->ops->release() frees the
> memory it could/should also reset the pointer to NULL

No it can't because it is not a pointer, it is contained by the backend 
specific structure.

eg.

struct dtpm_cpu {
	struct dtpm dtpm;
};

the release frees a dtpm_cpu structure.



> And if that is already done, the kfree below is harmless and there
> would be nothing to "fix".
>
> > kfree(dtpm);
> >
> > ?
>
> Kind regards
> Uffe


--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog